I’m not saying it’s “easy” to build a motor from scratch, but… if you decide you just need to build one, the Lebowski Axial is probably the “easiest” to build, I’ve been researching these things for a while now. Most motors are “radial flux”, like one cup spinning inside another slightly larger cup. However, there are some axial flux motors available for purchase, and the shafts are configured like one plate spinning next to the other (as shown in the picture here).
If you’re interested in this, I recommend first taking a quick look at the pictures in our article “Electric Motor Technology, Understanding the Terminology”. Many of the common questions you might have about motors will be answered in that article, so I’m writing this as if you’ve already read it.
I believe internally turning radial bearings are more popular in non-hub motor applications as they are more easily cooled passively through the stator with external aluminum fins attached. Outrunning hub motors can have ferrofluid added to significantly enhance passive cooling. If you can increase heat dissipation in the motor design, it means you can temporarily use high current for acceleration. For a while, Zero motorcycles used an early Motenergy axial-flux design, which had a rotor in the center and two stators, one on each side. Adding a fan or liquid cooling may help, but placing the thermal coils near the airflow around the vehicle helps with passive cooling.
The downside of axial flux is that to make the motor more powerful you can keep adding stators and rotors of the same diameter (which makes the motor wider), but… a stator nested in the center of the motor has a hard Passive cooling. Once engineers add liquid cooling to the equation to move heat away from the core, the added complexity, weight and cost could make other designs more desirable.
My interest in non-hub motor DIY motors stems from my fear of a possible future trade embargo affecting products from China. To be fair, you’ll still need to buy neodymium magnets and enameled magnet wire to build it, both of which are likely made in China.
________________________________________________
Who the hell is Lebowski?
If you think designing a motor that can be built in a DIY garage is impressive, he designed this just to help him design a DIY sine wave “Field Oriented Control” FOC controller! (link below). The name “Lebowski” was an avatar for a user on the e-bike chat forum Endless-sphere.com, and I’m grateful he posted a photo of this version there (Click here for a link to its motor section), a link to his discussion of this motor build is also at the bottom of this post.
Lebowski lives in Zurich, Switzerland, and here’s a short video showing how his DIY controller and motor work. His YouTube channel is “81FXB”
Configured as a three-phase “Permanent Magnet DC”/PMDC motor with an axial flux layout. His first prototype was a 250W single rotor/single stator version for a proof of concept. Click here to view this content. The prototype was able to deliver temporary peaks of 1000W without overheating.
The 2300W version shown here (60V x 38A = 2300 Watts) has three stators and four magnet rotors, two of which are double-sided.
________________________________________________
hub and magnetic rotor
For the spindle, he chose a solid bicycle rear axle hub that uses a “cassette freewheel” rather than the more common threaded freewheel. Manufacturers could keep adding rotors and stators, making this motor wider for more power potential, but… that would require a custom spindle.
The aluminum plate shown is the adapter that connects the spoke holes of the driver body to the rotor frame.
When there is some steel behind a permanent magnet, it attracts the magnetic flux on that side by providing a magnetic path, which causes the field on the other side of the magnet to become stronger and projected farther outward. Since magnets are expensive and steel backings are cheap, most PMDC motors have an “iron backing” behind the magnets. Steel is approximately 99.7% iron with a small amount of carbon added.
Of the four rotors in this design, the two outer rotors have magnets on only one side (facing the stator, obviously), and the two stators near the center of the motor have magnets on both sides, with the steel plate between them. The magnet holding frame is a normal aluminum frame.
Viewed from the side of the motor, the #one and #four rotors have magnets on only one side. The rotors #two and #third have magnets on either side, with the “back iron” sandwiched in between.
The rotor can be made lighter by drilling lots of holes into its frame, but here the aluminum mass provides the mass to act as a heat sink to help the magnets avoid overheating during acceleration thermal peaks (when the current is at its highest). Some common grades of neodymium magnets may begin to permanently lose some of their magnetism if the temperature reaches 178°F/80°C.
If the motor is cold under all conditions, it “could” be bigger, heavier and more expensive than necessary. If the temperature is too high, many things can be damaged, such as hall sensors, bearing seals, wire insulation and magnetic flux. Many of these can be damaged when approaching 200F/93C, I think… 140F/60C is a useful target for system design.
Samarium cobalt magnets were invented to be able to withstand very high temperatures, but…they are expensive.
Test install the first rotor. The face of the magnet is marked red and blue to designate north and south magnetic poles, the poles must alternate. You can see here that the four rotors will fit into the rotating part of the motor. There are ten magnets on each rotor face.
________________________________________________
stator
As we covered in our article on the history of the electrical term (click here to read it), the modern electromagnet was discovered by William Sturgeon, a key invention that made practical electric motors possible because electromagnets could switch on and off.
In the rotor shown above, the magnets are “permanent magnets”. In the stator, however, we see that the coils of copper wire can be turned on and off by the controller as needed at the precise moment when their flux causes the rotor to spin and perform work.
Lebowski chose round coils because they are easy to wind tightly, resulting in good “copper fill” in the available coil volume. The above device is easy to wrap them between the wooden discs. Whether you’re making a DIY wind turbine or a PMDC motor, the solenoid coils should be the same size and shape as the permanent magnets.I have seen DIY Motors and Generators with Rectangular Coils/Magnetsand “wedge” for maximum power density. Click here to see an example.
In the picture above, the coils are wound inside two discs of plastic film so the wood won’t stick when the coils are epoxyed into the stator (this will make sense in a minute).
Pictured above is the acrylic sheet used as one of the stator frames. The components of the rotating rotor assembly are mounted on the main shaft in the center and all stator assemblies are mounted on the outer corners.
I’ve seen single phase motors that hum a bit when all coils are energized and de-energized at the same time. I’ve also seen 5-phase motors, which I can say are smoother and more efficient, but…they’re also more complex, and they require a 5-phase controller, which isn’t common at all.
Combining the coils into three evenly spaced groups is the simplest configuration and is widely considered to be fairly smooth and quiet. Also, if you build 3-phase motors, you can use common and affordable controllers and “rotor position sensors”. In this design, Lebowski used nine coils in groups of three. The three coils shown below will be connected together to energize and de-energize simultaneously.
These are the three coils epoxy encapsulated into the stator frame. The copper wire looks just like normal bare wire, but the “magnet wire” is coated with a thin layer of flexible clear epoxy to electrically insulate one package from the next. Since the permanent magnet will pass through the stator frame, it must be made of a material that will not be affected by “eddy currents”, which cause drag and waste heat.
Lebowski opted for acrylic panels, I’ve also seen polycarbonate used. Both are fairly heat resistant.
Once each coil is inserted into place, its “pockets” are filled with clear hard epoxy. If this is not done, the coil vibrates causing the insulation coating of the wire to wear off, the coil shorts out, and the current takes a short path instead of flowing through each turn of the coil. The coils in this design are “coreless”, which means they do not have iron/steel in their center. Adding an iron core has advantages, but also disadvantages. Lebowski chose to go coreless for a number of reasons.
The AaABbBCcC coil name is only meaningful for how the coil is wired, as detailed in the build article. Additionally, the coil can be wound clockwise or counterclockwise, details that are critical to ensuring the coil is wound correctly.
There are ten magnets per rotor face and nine coils per stator. 250W version uses 14 magnets, 12 coils (see header image)
Note that the bottom rotor has one magnet layer, and the magnet layer above the stator is double-sided. The design uses a “single stator/twin rotor” design and stacks three for increased power potential.
The rightmost and leftmost outer rotors have only one layer of magnets (with a steel backplate), and the two inner rotors have two opposing layers of magnets with steel plates in between. The black cylinder on the left is the splined freehub for the sprocket, so this motor has a built in freehub.
Here’s another angle. One of the advantages of a freehub is that the sprockets can be easily moved in and out to help with chain alignment. As shown, the motor uses 27 coils and 60 magnets.
________________________________________________
boring technical details
“…it will be connected in series with the stator plate in WYE. Putting them in parallel would be terrible… acrylic glass can withstand very high temperatures. I put solder drops on it
There was an open flame touching it for short periods of time (for something with wire shrinkage), but that didn’t affect it at all. At 1000rpm it should deliver about 2000W mechanical output…no cogging. It runs as smooth and easy as a bike wheel…I run this motor with a sine wave controller and it runs flawlessly. The motor has a fair amount of inductance, about 20uH per coil if I remember correctly, and 9 coils in series per phase…from 1500 rpm and 2.3 kW mechanical power, the internet says it has a torque of 14.6 Nm…”
As a final note, you can use this configuration to make a single stator motor with two rotors that draws about 800W at 60V, or… a dual stator four rotor motor that draws about 1600W at 60V.
If you add a steel core to the coils, the resulting motor is more power-dense, but also creates a drag called “cogging” and also generates more waste heat.Zero Motorcycles used to do this for a while and used the Single double-sided center rotor with outer stators on both sides for better air coolingsuch as the Zero/Motenergy ME0913 motor. Below are detailed pictures of the single stator/single rotor version.
The full-size version shown in this article (three stators/six rotors, four of which are double-sided so it looks like four rotors) delivers 2300W at 60V, which is 52V/ It is also about 2000W under 14S.
________________________________________________
Link
2300W Motor Build Discussion (Click here)
250W/1000W “Peak” Prototype Discussion (Click here)
Riding for a year on a 250W prototype (Click here)
Lebowski’s YouTube channel (Click here)
Lebowski’s DIY Sine Wave FOC Controller Discussion (Click here)
________________________________________________
By Ron/spinningmagnets, June 2023
About Author
